When creating or modify databases and their tables, it is essential that you understand how the tables relate, or how you want them to relate. Normalization is the process whereby tables are optimised via referential integrity to remove duplication, redundancy, and errors.
Two key ways to represent the fields, tables, and relationships is via ER (Entity-Relationship) Diagrams and Data Dictionaries. The ER Diagram is just that; a diagram (similar to a UML Class or Object diagram) which shows the tables and their data fields, as well as the relationship between tables.
The Data Dictionary is listing of the formatting used to create you tables - eg relationships are viewed as Primary and Foreign keys etc. All the tables and their fields and formatting, as well as the Table Relationshops are also fully documented.
The ER Diagram is useful for taking a High Level view of the overall database including Relationships, where as the Data Dictionary is Essentially as it gives you key specific data - such as the field type, size, length, etc.
